methods for isolation of entomopathogenic fungi from the soil

Meyling 2007: Manual for isolation of soil borne entomopathogenic fungi 2

Methods for isolation of entomopathogenic fungi from the soil environment

Laboratory manual, January 2007

Deliverable 5.1. of VegQure under the research programme FØJO III

by Nicolai V. Meyling

Department of Ecology, Faculty of Life Sciences, University of Copenhagen,

Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark. [email protected]

Summary

Descriptions of methods and recommendation of laboratory procedures for the isolation of

soil borne entomopathogenic fungi (specifically Beauveria spp. and M. anisopliae) are

presented. For screening of occurrences of indigenous populations of entomopathogenic fungi

the insect bait method is recommended. Further recommendations are: 1) Collect sufficient

number of soil samples to cover the area of investigation; 2) if the bait method is used, apply

sufficient individuals of bait insects to each sample to increase the likelihood of isolating the

fungi present. Descriptions of isolation methods, statistical analyses of the data and

preparation of media and bait insects are given.


Introduction

Entomopathogenic fungi are natural enemies of insects and arachnids and the fungi contribute

to the regulation of their host populations. In agriculture, the fungi have been observed to

cause mortality in pest populations and several fungal species have been investigated for their

potential as biological control agents. The traditional approach in biological control with

entomopathogenic fungi has been to apply the fungal material (usually conidia) to the

cropping system, using an inundative or inoculative biological control strategy (Eilenberg et

al., 2001). Thus these approaches do not exploit the indigenous reservoir of fungi that is

already present in the cropping system.

Another biological control strategy is conservation biological control. Eilenberg et al.

(2001) defined this strategy as: "Modifications of the environment or existing practices to

protect and enhance specific natural enemies or other organisms to reduce the effects of

pests". Therefore knowledge of the community of natural enemies in the agroecosystem as

well as the effect of the agronomical practices on these organisms is essential to use a

conservation biological control strategy. This manual will focus on methods to obtain

knowledge of the community of entomopathogenic fungi in soils. The manual will be limited

to selected taxa of fungi belonging to the order Hypocreales in the division Ascomycota.

Entomopathogenic fungi occur naturally as infections in insect or arachnid hosts. Thus

sampling of host individuals can reveal information about host range and prevalence of fungal

species as pathogens in natural host population. However, several entomopathogenic fungi

only occur as infections in living hosts for a relatively short period of time during their life

cycle. The remainder of the life cycle these species presumably lurk as dormant conidia in the

soil in the vicinity of the dead host cadaver. Limited saprobic growth is some times possible

from resources contained in the host cadaver. Most fungi from the order Hypocreales are only

known in their anamorphic (asexual) life cycle in Europe, thus only mitosporic conidia are

formed. The dead host cadavers will mostly fall to the ground and thus a reservoir of fungal

material is present in the soil environment. Further dispersal from cadavers as focal points

presumably occur due to weather (rain and wind), soil manipulation and also insect activity

(Meyling et al., 2006).

Conidia produced on the surface of dead host cadavers are relatively long lived. These

structures represent the freeliving infective stages, as defined by Anderson & May (1981), of

the pathogen and need to come in contact with a susceptible host in order to grow and


proliferate successfully. In the laboratory, however, the conidia from hypocrealean

entomopathogenic fungi can also germinate, grow and conidiate in vitro on artificial rich

media. These two methods of germination are manipulated for isolation of entomopathogenic

fungi from the soil environment.

Methods for isolation of entomopathogenic fungi from soil samples

Selective media

A wide range of fungi occur in the soil environment and they have various ecological

functions. Most of these fungi, along with a range of bacteria, can grow on artificial media in

vitro. These abilities have long been exploited to isolate microorganisms from soil samples

and specific media have been developed to select for certain groups of microorganisms. Some

media for the selective isolation of entomopathogenic fungi have also been developed.

Bacteria can be inhibited by the application of broad-spectrum antibiotics such as

chloramphenicol, tetracycline or streptomycin (Goettel & Inglis, 1997). The main remaining

obstacle in using this isolation method is that the hypocrealean entomopathogenic fungi grow

relatively slowly in comparison to the ubiquitous opportunistic saprotrophic fungi found in

the soil environment. Thus the contents of the media need to include substances that prevent

these fungi from overgrowing the species of interest. Generally, the species Metarhizium

anisopliae, Beauveria bassiana and B. brongniartii have been investigated the most.

Media for isolation of Metarhizium spp.

Goettel & Inglis (1997) provide a list of suitable selective media for Beauveria and

Metarhizium (Goettel & Inglis, 1997, p. 248). The suggested medium for isolation of

Metarhizium spp. is often called Veens semiselective medium (Hu & St Leger, 2002) to refer

to its first description by Veen & Ferron (1966). The medium contains the antibiotics

chloramphenicol as well as the fungicides dodine and cyclohexamide (Goettel & Inglis,

1997). In different laboratories modifications have usually been made to optimise isolation

results based on experience. For example, Hu & St. Leger (2002) used Veens medium to

isolate M. anispoliae, but omitted cyclohexamide to study the occurrences of other fungi than

M. anisopliae. In our laboratory, a modified medium has been developed to re-isolate applied


conidia of M. anisopliae in order to estimate persistence in the soil as well as vertical

movement (Vestergaard & Eilenberg, 2000). The procedure to make this medium is described

in Appendix A.

Media for isolation of Beauveria spp.

During the last decades research groups have been developing biocontrol programmes for the

control of soil dwelling larvae of scarabaeid beetles, principally the cockchafer, Melolontha

melolontha. This group of beetles is frequently infected in the field by the pathogen B.

brongniartii and this particular fungal species has been developed as a biocontrol agent in

Switzerland and Austria. In order to monitor the fate of applied fungal material in the soil, a

selective medium was developed. Originally described by Strasser et al. (1996) this medium

has been used in several studies (Enkerli et al., 2004; Keller et al., 2003; Kessler et al., 2003;

Kessler et al., 2004). At KVL, this medium is also used for the detection of survival of

applied B. brongniartii material, but has also been successfully used to isolate B. bassiana

from phylloplanes of different plant species (Meyling & Eilenberg, 2006a). This medium is

described in Appendix A.

Insect bait method

The use of selective media exploits the saprotrophic abilities of hypocrealean

entomopathogenic fungi. However, to exploit the ability of the fungi to infect host, the insect

bait method can be used. This method was originally developed to isolate entomopathogenic

nematodes from soil samples, but fungi were sometimes additionally isolated (Zimmermann,

1986). Thus Zimmermann (1986) suggested that this method could also be a standard

isolation method for entomopathogenic fungi. For the method to be feasible insects, which are

easily reared and are susceptible to the fungi, must be used. The traditional bait insect is the

highly susceptible larvae of the wax moth, Galleria mellonella, (Lepidoptera: Pyralidae) but

also mealworm larvae, Tenebrio molitor (Coleoptera: Tenebrionidae), are suitable. Baiting

soil samples with larvae of G. mellonella is a widely applied tool to screen for indigenous

species of entomopathogenic fungi (Vanninen et al., 1989; Vänninen, 1996; Chandler et al.,

1997; Bidochka et al., 1998; Klingen et al., 2002; Keller et al., 2003; Meyling & Eilenberg,

2006b). Method for rearing G. mellonella is presented in Appendix C.


Few studies have evaluated the use of several bait insects from different taxa. Klingen et al.

(2002) found that dipteran larvae isolated fungi differently than G. mellonella. More

specifically, larvae of Delia floralis (family Anthomyiidae) isolated Tolypocladium

cylindrosporum more frequently than did G. mellonella (Klingen et al., 2002). Thus the use of

insect baits can also be considered to be a selective isolation method. However, the "Galleria

bait method" appears to be more sensitive than traditional plating on media (Keller et al.,

2003) and is therefore useful for isolation and identification of the spectrum of

entomopathogenic fungi indigenously present in soils.

Recommendations for the use of the insect bait method

Since Zimmermann (1986) recommended the insect bait method for the selective isolation of

entomopathogenic fungi, numerous studies have been carried out using insect baits, especially

G. mellonella. In 1998, a further set of recommendations was published by the International

Organisation for Biological Control (IOBC) (Zimmerman, 1998). Here, the recommendations

are:

• Air-dry soil samples and re-moisturise the samples afterwards to appropriate levels to

avoid infections by entomopathogenic nematodes

• Use 5-10 larvae per sample

• Replicate baiting of each sample 5 times

• Incubate the samples at 20-22oC in the dark and invert the individual containers every

day during the first week

• Inspect the samples for the first time after 5 days and repeat this every 3-4 days for 3

weeks after initial baiting

• Surface sterilise the dead bait larvae with 1% Na-Hypochlorite prior to incubation in

moist chamber


Based on my own experience, the following comments are given to the usage of the insect

bait method to screen soil samples for entomopathogenic fungi. These comments are

supplementary to the recommendations given by Zimmermann above.

• There is no need to first dry the sample and the re-moisture it. However, it takes some

experience to get a feeling of what the best moist contents of the soil should be. As a

rough guideline, the soil should not be too damp and not leave too much condensation

on the inside of the container. However, some condensation on the inside of containers

is desirable. The soil should not be so moist that clumps are formed. Remember to

punch air holes in the lids of containers

• It is important to turn the containers regularly in the beginning of the baiting period

(first week) to make bait insects penetrate as much soil as possible while they are still

vigorous

• If G. mellonella larvae are used, select medium sized larvae and prepare them by heat

treatment in warm water to prevent extensive webbing in the soil. The method was

used successfully by Meyling & Eilenberg (2006b) and is described in Appendix 2

• Use 10 bait larvae as some always disappear or die of causes other than mycosis

• There is no need to inspect the samples until after 1 week because no larvae die of

mycosis during the first 5 days at 20-22oC

• Replicating baiting of each sample is fine if the number of samples is low. Otherwise,

do the replication in the field and take more samples. Then there is no problem of

relatedness of the results during statistical analysis

• Surface sterilisation is fine to prevent external saprophytic fungi from growing on the

dead cadaver. However, if the larvae are indeed killed by entomopathogenic fungi that

have penetrated the body of the insect they will immediately emerge from the cadaver

keeping other opportunistic fungi at bay. Furthermore, individual surface sterilisation

of large numbers of larvae will be a huge amount of work that does not provide much

information if many soil samples are to be screened. Thus surface sterilisation should

be considered critically and evaluated with regards to the number of samples in the

study


Advantages and disadvantages of using different isolation methods

Each isolation method will have some advantages and disadvantages. These have to be kept in

mind while selecting the most suitable method to be used for a specific study.

Soil suspensions on selective media

Advantages • Quantitative data

• Parametric data; use of standard statistical analyses (e.g. ANOVA)

Disadvantages • Overgrowth of opportunistic soil fungi on media

• Small soil sample (1 g); risk of not sampling the fungus because

entomopathogenic fungi are usually clumped in the soil

• Dilution effects: zero-values when in fact the fungus is present

because only a diluted sample is taken from 1 g sample of the

original sample

Insect bait method

Advantages • Use of G. mellonella is a very sensitive detection method

• Entomopathogenic fungi are selectively isolated

Disadvantages • Some insect species may select for specific fungal pathogens

• Moist soil may enhance the infection of nematodes and not fungi

• Difficult to quantify inoculum levels

Soil sampling, types of data and their analysis

Traditional analyses

Any investigation of the occurrence of entomopathogenic fungi in soil needs to consider

appropriate methods for statistical analysis of the data. The most widespread and always

applicable way to analyse occurrence data is by the use of frequencies (qualitative data) and

chi-square tests. No considerations about distribution of data are necessary for this type of

test. Since frequency based data only inform about +/- occurrence it is essential that a


sufficient number of samples are included in the analysis and that an appropriate effort for

isolation of fungi from each soil sample is applied. Chi-square test needs to have at least

numbers larger than 5 in each cell (or 20% of cells) to be reliable thus the number of positive

samples needs to exceed 5. For instance, if the "true" occurrence of a fungus is approximately

10%, then at least 50 samples need to be included for isolation. If the bait technique is used

then enough individuals of bait insects need to be applied to each sample to yield reliable

data. For example, Chandler et al. (1997) only used one G. mellonella larva for each sample

and found low frequencies compared to other studies. This result is probably due to the death

of larvae of other causes than fungal infections thus presumably underestimating the

occurrence of fungi.

When the soil dilution plating method is used the procedure is more tedious than the

bait method (preparation of soil subsamples, determination of water content, dilution and

plating of suspensions as well as the initial production of media). Thus lower numbers of

samples are usually included than is possible if the insect bait method is used. The dilution

plating method yields quantitative data that can readily be analysed by parametric methods

(e.g. ANOVA) normally after transformation of the data to stabilise variances.

Distribution patterns

Most published studies have reported on collection of soil samples with no particular

references to the spatial distribution of the individual samples. Normally, the samples have

been described with respect to the type of habitat from which they were collected, e.g.

agricultural field soil, forest soil, etc. Meyling & Eilenberg (2006b) collected soil samples

from specific points in a sampling grid based on GIS (Geographical Information Systems).

The individual sampling points could be identified by GPS (Global Positioning System). Thus

the occurrence of entomopathogenic fungi in each sample could be related to a specific

coordinate and a map of the occurrences could be created. The occurrence data were analysed

as quantitative data since the number of dead larvae (0 - 10) was included in the analyses.

Spatial statistics, where both data values as well as locations of the data in two-dimensional

space are included, were applied to the data. The method used is called SADIE and has been

developed for analysis of count data such as the occurrence of insects in traps etc, but the

method was also found to be useful for data on occurrence of entomopathogenic fungi. More

information of the statistical method can be found on the website of Professor Joe Perry:


http://www.rothamsted.bbsrc.ac.uk/pie/sadie/SADIE_home_page_1.htm and in Perry et al.

(1999).

The main advantage using the method is that it allows for the identification of clusters of

patches and gaps of the organisms and makes explicit tests of whether the data follow a

random distribution or are clumped. In the study by Meyling & Eilenberg (2006b) the fungus

Beauveria bassiana was found to be distributed in clumps in one year and clusters of patches

and gaps of the fungus were found in specific parts of the investigated field (Figure 1).

Furthermore, sampling of soil in selected patch and gap areas and subsequent isolation of

entomopathogenic fungi confirmed the identification of the areas as areas of high and low

occurrence of B. bassiana (Meyling & Eilenberg, 2006b; Figure 1). These explicit results

could not have been obtained if knowledge of the location of each sampling point had not

been available. Such data provide the opportunity to correlate occurrences to other spatial

factors in the cropping system and subsequently develop hypotheses about factors that could

effect the distribution of entomopathogenic fungi. For example, combination of data of

selected insect populations (hosts) and fungal inoculum could provide results of the

correlations in distribution between the populations of these organisms.


Figure 1

Horizontal distribution of the entomopathogenic fungus B. bassiana in the field at Bakkegården in

September 2002 calculated by the statistical software SADIE. The scale of the figure is in metres.

Clustering indices equal to 0 are represented by contour lines accompanied by 0. Grey shaded areas

represent ‘gaps’ i.e. areas where clustering indices are below –1.5 (areas with lesser occurrence of B.

bassiana than what should be expected from a random distribution). Black areas are ‘patches’, i.e.

where clustering indices are above 1.5 (areas with higher occurrence of B. bassiana than what should

be expected from a random distribution). The two small squares enclose the area of sampling with

reduced distances conducted in September 2003 and these data confirmed the existence of patches

and gaps. Data are presented in Meyling & Eilenberg (2006b).

0 50 100 150 200 250 300 350 400 450 5000

50

100

150

200

250

300

350

400

450

500

550

N


References

Anderson, R.M. & May, R.M. (1981) The population dynamics of microparasites and their

invertebrate hosts. Philosophical Transactions of the Royal Society of London Series B

- Biological Sciences, 291, 451-524.

Bidochka, M.J., Kasperski, J.E. & Wild, G.A.M. (1998) Occurrence of the entomopathogenic

fungi Metarhizium anisopliae and Beauveria bassiana in soils from temperate and

near- northern habitats. Canadian Journal of Botany, 76, 1198-1204.

Chandler, D., Hay, D. & Reid, A.P. (1997) Sampling and occurrence of entomopathogenic

fungi and nematodes in UK soils. Applied Soil Ecology, 5, 133-141.

Eilenberg, J., Hajek, A. & Lomer, C. (2001) Suggestions for unifying the terminology in

biological control. Biocontrol, 46, 387-400.

Enkerli, J., Widmer, F. & Keller, S. (2004) Long-term field persistence of Beauveria

brongniartii strains applied as biocontrol agents against European cockchafer larvae in

Switzerland. Biological Control, 29, 115-123.

Goettel, M.S. & Inglis, G.D. (1997) Fungi: Hyphomycetes. Manual of Techniques in Insect

Pathology (ed. by L.A. Lacey), pp. 213-249. Academic Press, San Diego, USA.

Hu, G. & St. Leger, J. (2002) Field studies using a recombinant mycoinsecticide

(Metarhizium anisopliae) reveal that it is rhizosphere competent. Applied and

Environmental Microbiology, 68, 6383-6387.

Keller, S., Kessler, P. & Schweizer, C. (2003) Distribution of insect pathogenic soil fungi in

Switzerland with special reference to Beauveria brongniartii and Metharhizium

anisopliae. Biocontrol, 48, 307-319.

Kessler, P., Enkerli, J., Schweizer, C. & Keller, S. (2004) Survival of Beauveria brongniartii

in the soil after application as a biocontrol agent against the European cockchafer

Melolontha melolontha. Biocontrol, 49, 563-581.


Kessler, P., Matzke, H. & Keller, S. (2003) The effect of application time and soil factors on

the occurrence of Beauveria brongniartii applied as a biological control agent in soil.

Journal of Invertebrate Pathology, 84, 15-23.

Klingen, I., Eilenberg, J. & Meadow, R. (2002) Effects of farming system, field margins and

bait insect on the occurrence of insect pathogenic fungi in soils. Agriculture

Ecosystems & Environment, 91, 191-198.

Meyling, N.V. & Eilenberg, J. (2006a) Isolation and characterisation of Beauveria bassiana

isolates from phylloplanes of hedgerow vegetation. Mycological Research, 110, 188-

195.

Meyling, N.V. & Eilenberg, J. (2006b) Occurrence and distribution of soil borne

entomopathogenic fungi within a single organic agroecosystem. Agriculture

Ecosystems & Environment, 113, 336-341.

Meyling, N.V., Pell, J.K. & Eilenberg, J. (2006) Dispersal of Beauveria bassiana by the

activity of nettle insects. Journal of Invertebrate Pathology, 93, 121-126.

Perry, J.N., Winder, L., Holland, J.M. & Alston, R.D. (1999) Red-blue plots for detecting

clusters in count data. Ecology Letters, 2, 106-113.

Strasser, H., Forer, A. & Schinner, F. (1996) Development of media for the selective isolation

and maintenance of virulence of Beauveria brongniartii. Microbial Control of Soil

Dwelling Pests. AgResearch, Lincoln, New Zealand (ed. by T.A. Jackson and T.R.

Glare), pp. 125-130.

Vänninen, I. (1996) Distribution and occurrence of four entomopathogenic fungi in Finland:

Effect of geographical location, habitat type and soil type. Mycological Research, 100,

93-101.

Vanninen, I., Husberg, G.B. & Hokkanen, M.T. (1989) Occurrence of entomopathogenic

fungi and entomoparasitic nematodes in cultivated soils in Finland. Acta

Entomologica Fennica, 53.


Veen, K.H. & Ferron, P. (1966) A selective medium for isolation of Beauveria tenella and of

Metarrhizium anisopliae. Journal of Invertebrate Pathology, 8, 268-269.

Vestergaard, S. & Eilenberg, J. (2000) Persistence of released Metarhizium anisopliae in soil

and prevalence in ground and rove beetles. IOBC/WPRS Working group: Insect

pathogens and insect parasitic nematodes, Vienna, Austria, March 22-26, 1999, 23,

181-185.

Zimmerman, G. (1998) Suggestions for a standardised method for reisolation of

entomopathogenic fungi from soil using the bait method (G. Zimmermann, J. Appl.

Ent. 102,213-215, 1986). IOBC/WPRS Bulletin, Insect pathogens and insect parasitic

nematodes, 21, 289.

Zimmermann, G. (1986) The Galleria bait method for detection of entomopathogenic fungi in

soil. Journal of Applied Entomology, 102, 213-215.


Appendix A: Selective media for isolation of entomopathogenic fungi

SM

Selective medium

• Suspend 32.5 gram SDA (Sabouraud Dextrose Agar) in 500 ml distilled water in a

blue cap bottle.

• Add 1 ml Dodine (a fungicide to inhibit fungal growth)

• Mix the medium and mark the blue cap bottle with autoclave tape.

• Autoclave the medium for 20 min at 120 °C 20 bar.

• (Remember that the lid of the blue cap bottle has to be loose during autoclaving)

• Cool the medium after autoclaving to approx. 60°C and add:

• 500 µl Chloramphenicol (antibiotic, inhibits bacteria)

• 500 µl Streptomycin sulphate (antibiotic, inhibits bacteria)

• Invert the bottle gently and pour the plates.

Solutions:

Chloramphenicol: 1g in 10 ml 96% ethanol.

Streptomycin sulphate: 0,5 g in 10 ml sterilized distilled water.

Dodine: 5 g in 45 ml distilled water.


Appendix A: Selective media for isolation of entomopathogenic fungi

BSM

Selective medium for Beauveria spp.

5 g Peptone Dissolve in 500 ml dem. water

10 g Glucose pH adjusted to 6.3 with 1 M HCl

6 g Agar ( no.1, Oxoid) Autoclave for 20 min. at 120 oC

When the medium has cooled to 50-60 o C add:

0.5 ml a’ 0.6 g/ml Streptomycin

0.5 ml a’ 0.05 g/ml Tetracycline

0.5 ml a’ 0.1 g/ml Dodine

2.5 ml a’ 0.05 g/5 ml Cyclohexamide

Invert gently the bottle without making air bobbles and pour the plates


Appendix B

Treatment of Galleria mellonella larvae to prevent webbing in soil

When G. mellonella are reared at 20oC, four week old larvae are most suited for baiting to

avoid that they pupate in the soil. To prepare the heat treatment:

1. Place a beaker with 500 ml of water in a water bath at 56oC.

2. Take the number of larvae (+10%) from the rearing containers and place them in a

box. Place a sheet of paper in the box and the larvae will crawl under this for hiding.

Thus they do not crawl out.

3. Remove the paper and shake the box so that the larvae can not cling to webbing in the

box. Pour all the larvae into the beaker with hot water. Let them remain in the water

for 10 seconds, maximum 15 s. Pour the water through a sieve and cool the larvae in

cold running water for 30 s. Place the larvae on dry tissue paper and place them in the

dark for 3-5 hours.

4. When the larvae have recovered from the treatment (they may appear dead at first)

place them in the containers with soil. Do not invert the samples until the following

day as the larvae may be squashed and die.

This description is based on recommendation from Ingeborg Klingen, Norway, and

Woodring, J. L. & Kaya, H. K. 1988. Steinernematid and Heterorhabditid Nematodes: A

handbook of biology and techniques, Fayetteville, Arkansas: Arkansas Agricultural

Experiment Station. Pp. 1-30


Appendix C. Rearing of the wax moth, Galleria mellonella

Rearing can be performed in plastic boxes incubated in the dark in a climate controlled room

at 20oC. Adult moths should be provided with a solution of water and honey. Under the lid of

the box containing the adults, strips of folded paper can be provided for oviposition. The

females will attempt to place their eggs in crevices as the folded paper represents. The paper

can then easily be removed with the eggs attached.

Paper strips with eggs can be placed in a new box with a ball of food for early instars (see

below). When the eggs hatch, the neonate larvae will themselves move to the food and start

feeding. When the larvae have reached approximately 1 cm in length they can be provided

with food for late instars (see below). Larvae of approximately 2.5-3 cm in length (4 weeks

after hatching) are suitable for baiting soil samples.

Food for early instar larvae 180 g honey 260 g whole grain wheat flour 180 g glycerine 80 g dry brewers yeast 50 g bee wax 50 g wheat bran Honey, glycerine and bee wax are melted in a cooking pot (don't boil). Remove from heat. Add brewers yeast and then whole grain wheat flour. Then add wheat bran. Mix thoroughly. Form the mixture into balls. These can be kept in the fridge until use. Food for late instar larvae 280 g honey 240 g glycerine 40 g dry brewers yeast 400 g blended dry dog food (e.g. Pedigree Junior) 100 g rolled oats 100 g wheat bran Mix honey, glycerine and brewers yeast and add blended dog food (must be blended to powder). Add rolled oats and wheat bran. If the mixture is too greasy add more oats and wheat bran. Keep the food refrigerated and add the food to boxes with late instar larvae.

methods for isolation of entomopathogenic fungi from the soil

Documents